CosyVoice

CosyVoice is an intelligent voice input application from Alibaba’s Tongyi for Windows. It converts speech to text in real time and is intended for people who want to create or organize content by speaking. During transcription, it can remove conversational fillers such as “um,” “ah,” and “that,” along with repeated words. It can also recognize spoken corrections such as “no, change it to…” and apply the revised wording automatically.

The application can turn unstructured dictation into numbered lists, tables, or outlines. A single spoken instruction can generate a complete email or meeting invitation. CosyVoice supports transcription of dialects including Shanghainese, Cantonese, and Sichuanese into standard Mandarin. It also provides multilingual translation, speaker distinction, intelligent summaries, and follow-up tracking for tasks.

CosyVoice fits the Windows software category as a voice-based writing and productivity tool.